Work things out with the AI assistant you already use. It investigates, proposes and questions; you judge and decide. It works from your own terms and rules, and each conversation picks up where the last one left off. You don't need to know how Kheel is laid out.
Kheel has no chatbot of its own: you keep the assistant you have, and it uses Kheel's tools to work in your model. It works with any assistant that supports MCP, such as Claude and GitHub Copilot.
We're introducing group appointments. What in our model might need to change?
Reading the modelFound 3 affected, 1 open question
An appointment is for one patient with one clinician today, so that definition, double booking and late cancellation are affected. Nothing says whether a group session is one booking or several linked ones. That needs deciding before the capability is written, so I've raised it as an open question.
We're building software for dental practices. Help me work out how appointments, bookings, clinicians and treatment plans relate.
Drafting the domainProposed 4 terms and how they relate
I've proposed separating an appointment, the clinical visit, from the booking that reserves its time and chair, and linked a treatment plan to the appointments it needs. Does that distinction reflect how your practices work?
Map the activities in booking a new patient, from their first call to a confirmed booking.
Mapping the activitiesMapped 6 activities
Mapped six activities between the patient and the front desk. Nobody said who checks insurance cover, so I've marked that step as an open question.
Write user stories for rescheduling an appointment, from the front desk's side, with acceptance criteria.
Writing user storiesAdded 3 stories
Added three stories to Scheduling, with their acceptance criteria. They're marked as drafts for you to review.
A prospect told me Molarix has cut its entry price to $49 a chair. Log it, and tell me if it changes our read.
Updating MolarixLogged a move, revised the read
Logged it as a pricing move. They're now cheaper than us for single-chair clinics, where we win most deals, so I've raised the threat to high.
Your assistant helps investigate the market, develop the domain model, define capabilities and prepare specifications. Say what you want, and it finds the right place in the model.
Investigate customers, competitors and the domain together. What you find becomes market and domain records.
“Here are my notes from today's call. Log anything they said about competitors.”
Clarify the concepts and how they relate, the jobs people need done and the rules that hold.
“Two clinics say drop-in and one says walk-in. Are they the same thing?”
Work through capabilities, use cases and acceptance criteria, each linked to the part of the domain it serves.
“Draft the stories for rescheduling, and leave them as drafts for me to review.”
Look for ambiguities, gaps and conflicting assumptions, with the model as something to reason against.
“Which personas have we never checked with real people?”
Work out design decisions and specifications, traced back to the capability and the need behind them.
“Why do rescheduled appointments have to stay in the same chair?”
Revisit assumptions and decisions as the product changes. The model carries across conversations, people and agents.
“Which competitors haven't we looked at in over a month?”
A skill gives an AI assistant instructions for a particular kind of task. Ours are free and work with or without a Kheel account; with Kheel connected, your assistant can keep what they work out as records.
Connecting takes a minute: add Kheel to your assistant, sign in and approve. Free for three months while Kheel is in beta.
Allow your AI assistant to use Kheel?